Biostatistics Weekly Seminar

Randomized trials are considered the gold standard for estimating causal effects. Trial findings are often used to inform health policy, yet their results may not generalize well to a relevant target population due to potential differences in effect moderators between the trial and population. Statistical methods have been developed to improve generalizability by supplementing trial data with non-experimental population data, yet identifying a suitable data source on the target population of interest can be challenging in practice. Large health surveys with complex survey designs are one logical source for population data; however, there is currently no best practice for properly accounting for the survey design when implementing generalization methods. In this talk, I propose an approach to estimating the population average treatment effect while incorporating population survey weights in this context. I then examine the performance of these methods (and the potential consequences of ignoring the complex survey design) through simulation, and apply the methods to generalize findings from a trial evaluating a web-based intervention for treating substance use disorders, to a target population from NSDUH. This work highlights the importance in properly accounting for the complex survey design when generalizing trial findings to a population represented by a complex survey sample.

Zoom (Link to Follow)
11 November 2020
1:30pm